Contact

Kubernetes Platform Engineer, GPU Clusters

You build and run the Kubernetes layer under Rillor's own platform, the private superintelligence (SI) environments we run for clients and the Kubernetes platforms we build for them. You turn bare-metal GPU servers into clusters that schedule GPUs well, upgrade cleanly and keep each tenant inside its own boundary.

Apply

Team: GPU cluster deployment, Kind: Full-time, Location: Remote, United States, Pay: $124,000 to $201,000 a year

Duties

  • Stand up Kubernetes on bare-metal GPU servers with kubeadm or Cluster API, from the first node to a highly available control plane.
  • Run the cluster lifecycle: version upgrades, node drains and replacements, certificate rotation, etcd backups and restores.
  • Enable GPUs with the NVIDIA device plugin or GPU Operator. Keep driver and CUDA versions under control, and export GPU health with DCGM.
  • Set up GPU scheduling for training and serving: labels and taints, quotas per tenant, MIG or time slicing where it fits, queueing with Kueue or Volcano.
  • Write operators and controllers in Go or Python where standard tools fall short.
  • Connect workloads to block, object and file storage through CSI drivers.
  • Give each tenant its own namespace with quotas and its own credentials, and keep that layout declared in Git.
  • Write a runbook for each failure you fix, so the client's team can run the platform after handover.

What you bring

  • 3+ years running Kubernetes in production, including at least one cluster you built yourself on servers or virtual machines.
  • You know how Kubernetes works inside: how the control plane, the node agents and custom controllers fit together.
  • Hands-on work with GPU nodes in Kubernetes, including device plugins, drivers and the NVIDIA container toolkit.
  • Strong Linux skills across systemd, networking basics, storage and driver troubleshooting.
  • Code in Go or Python, and comfort with Helm, Kustomize and Git-based workflows.
  • Cluster and GPU metrics with Prometheus and Grafana.

Nice to have

  • CKA or CKS.
  • Slurm on Kubernetes or another batch scheduler for training.
  • Topology-aware scheduling for NVLink and InfiniBand, or RDMA device setup for pods.
  • Rook or another operator that runs storage inside Kubernetes.
  • Open source work on Kubernetes or other CNCF projects.

Pay and benefits

$124,000 to $201,000 a year

Full-time W-2 employment with health insurance, a retirement plan and paid leave.

No application deadline. We accept applications for this role on an ongoing basis.

Location

Remote, United States

Apply

Apply

We review applications as each team grows and contact you when there is a match.

Rillor is an equal opportunity employer. We consider qualified applicants without regard to race, color, religion, sex, sexual orientation, gender identity, national origin, age, disability, veteran status or any other status protected by law.

If you need an accommodation to apply, email careers@rillor.com.

Notices

Trademarks. NVIDIA, AMD, Ceph, Kubernetes and related product names are trademarks of their owners and are used only to identify products. No affiliation or endorsement is implied.